影像式全自动检定中线纹尺噪声复原技术研究  

Research on noise restoration technology of linear rule in image automatic verification

作　　者：李淼 Li Miao(China Airborne Missile Academy,Luoyang Henan,471003)

机构地区：[1]中国空空导弹研究院计量测试中心

出　　处：《电子测试》2020年第5期52-54,34,共4页Electronic Test

摘　　要：随着现代计量测试仪器的创新发展,基于视觉检测技术和机电控制技术的影像式自动检定设备越来越多的被采用到计量检定工作中,由于线纹量具使用频率较高,难免引起小面积的线纹污损、线纹缺失,虽然大概率情况下并不影响使用,但在检定过程中,通常需要人工对无法识别的线纹进行手动选线,如此一来自动化检定装置则失去意义。本文以目前已投产的"影像式线纹尺检定仪"为项目开展基础,引入近年热点研究领域--人工智能,基于DCC算法进行图像模型聚类分析,基于WGAN算法提出图像修复模型。With the innovation and development of modern measurement and test instruments, image automatic verification equipment based on visual inspection technology and electromechanical control technology is more and more used in the measurement and verification work. Because of the high frequency of use of linear measuring tools, it is inevitable to cause small area of line mark damage and loss of line mark. Although it does not affect the use of large probability, but in the verification process, through It is often necessary to select lines manually for unrecognized lines, so that the automatic verification device will lose its significance. Based on the project "image linear ruler calibrator" which has been put into production, this paper introduces the hot research field in recent years-artificial intelligence, cluster analysis of image model based on DCC algorithm, and propose image restoration model based on wgan algorithm.

关 键 词：线纹尺自动检定系统 图像修复 聚类算法 无监督学习 训练模型 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术] TH711[自动化与计算机技术—计算机科学与技术]